Gene Epstein of Barron’s generally focuses on the official federal unemployment rate, U-3, even though that number omits a number of people who are not fully engaged in the work force despite their interest in finding a job. But Epstein explains in his latest “Economic Beat” column why he believes the official number is about a half-percentage point too low.

The BLS keeps six measures of labor underutilization, “U-1” through “U-6,” of which U-3 is the official measure. U-3 covers only those jobless folks 16 and older who have looked for work over the past four weeks. U-6 includes those folks and adds two other categories, often referred to as the “hidden unemployed.” The first is the “marginally attached”—people who haven’t looked for a job over the past four weeks, but have done so over the past 12 months. The second consists of the involuntary part-timers (“part-time for economic reasons,” in BLS parlance)—people who work part-time, but are searching for full-time positions.

Naturally, U-6 has always been a lot higher than U-3, prompting many compassionate observers into championing U-6 as the more accurate measure of the “true” rate of joblessness. But while there are certainly hidden unemployed not captured in U-3, I have generally focused on the official measure, so long as it tracks U-6 within a reasonable range.

The trouble is, that range has been exceeded. But meanwhile, consistent advocates of U-6 might ponder the following. For 15 months from October 1999 through December 2000, U-3 fluctuated between 3.8% and 4.1%—by all accounts, a time when jobs were quite plentiful and the labor markets unusually tight. Yet through this same 15 months, U-6 ran between 6.8% and 7.2%, averaging 177% higher. And it turns out that, over the 235 months since January 1994 when the BLS began tracking U-6, the ratio between U-6 and U-3 has also been 177%. Over that period, the ratio has fluctuated between a low of 163%, in ’02 and ’03, and a high of 189%. When the ratio gets that high, U-6 may be trying to tell us something.

That high of 189% was reached just last month. In July 2013, U-6 was at 14%, and if you assume a “normal” ratio last month of 177%, then U-3 would be 7.9%, not 7.4%. Also, if you parse U-6 you find that, the reason it’s unusually high is not because of the marginally attached, but because of the unusually high share of involuntary part-timers.

For now, at least, we should keep our eyes on both measures.